--- Day changed Mon Apr 14 2008
00:00 < Aulere> ok, I get phase locked loop
00:01 < Aulere> so the auditory wave is phase locked to the brain wave which then increases attention how?
00:01 < fenn> uh, the other way around
00:01 < fenn> the brain wave is locked onto the auditory signal
00:01 < Aulere> ok
00:02 < Aulere> right, sorry, that's what I meant.
00:02 < fenn> i wouldnt say it increases attention, rather it shifts attention
00:02 < Aulere> oh?
00:02 < fenn> yeah, which causes more attention, bach or eminem
00:03 < fenn> (trick question)
00:03 < Aulere> ok, so how would it do that? how would brain wave locking increase oxygen/glucose consumption in attentional areas?
00:04 < Aulere> in certain attentional areas and not others?
00:05 < fenn> er.. i dunno
00:05 < Aulere> ok no worries
00:06 < fenn> once you get into cell differentiation i'm pretty clueless
00:06 < Aulere> no problem. thanks for helping :)
00:07 < Aulere> where are you from?
00:09 < fenn> bloomington indiana
00:09 < fenn> its like austin texas, i hear
00:09 < Aulere> cool. I've been to Indiana
00:09 < fenn> university town, in a sea of conservativism
00:09 < Aulere> really? nah. Texas has huge bugs
00:10 < Aulere> ah
00:10 < Aulere> are you a student?
00:11 < Aulere> (whereas I am in a sea of liberalism - Northampton, MA)
00:12 < fenn> not sure if i'm a student or not. there's no money exchanging hands for my learning, if that's what you want to know
00:13 < Aulere> ? 
00:13 < Aulere> mine niether actually. But I'm definitely a student.
00:13 < fenn> i'm in search of an ethically consistent financial strategy
00:13 < Aulere> *neither
00:13 < Aulere> hehe
00:13 < Aulere> sounds interesting
00:14 < fenn> you mean, 'sounds like an impossible problem youve cooked up'
00:14 < Aulere> heh
00:14 < fenn> dont worry, its just NP complete
00:14 < Aulere> NP?
00:15 < fenn> nevermind
00:16 < Aulere> nondeterministic polynomial time?
00:17 < Aulere> by the by, I'm on #ai to learn; not because I know much about the more esoteric aspects, yet.
00:17 < Aulere> *"esoteric" (comic effect again)
00:17 < Aulere> so what do you study? or where do you work?
00:20 < Aulere> computers.
00:24 < fenn> ostensibly i'm following a technological development path that should lead to my very own space colony in twenty years or so, but i'm interested in open hardware as a movement in the meanwhile
00:24 < Aulere> cool
00:25 < Aulere> Well, time for bed for me. Night.
00:26 -!- Aulere [n=dragon_d@131.229.176.252] has quit []
02:28 -!- facefaceface [n=chatzill@bioinformatics.org] has joined #hplusroadmap
02:28 -!- facefaceface is now known as faceface
07:19 -!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] has joined #hplusroadmap
07:25 < kanzure> krebs: !help
07:25 < kanzure> !help
07:25 < krebs> help topics: 6 core modules: auth, basics, config, httputil, remote, userdata; 73 plugins: alias, autoop, autorejoin, azgame, babel, bans, bash, cal, chanserv, chucknorris, debug, deepthoughts, delicious, dice, dict, digg, dns, eightball, excuse, factoids, figlet, forecast, fortune, freshmeat, grouphug, hl2, host, imdb, insult, iplookup, karma, keywords, lart, lastfm, linkbot, markov, math, modes, nickserv, q, quiz, quote, reaction, realm, remind,
07:25 < krebs> remotectl, ri, roshambo, rot, roulette, rss, salut, script, search, seen, shiritori, shortenurls, slashdot, spell, theyfightcrime, threat, time, topic, translator, tube, twitter, unicode, urban, url, usermodes, weather, wheeloffortune, wserver (help <topic> for more info)
07:38 < kanzure> hosting DTDs looks like a viable option
07:39 < kanzure> it's a good way to associate/aggregate metadata about a file format all at one place
07:42 < kanzure> what type of bullshit is this? "Study calls free will into question"
07:42 < kanzure> http://science.slashdot.org/article.pl?sid=08/04/13/2052206
07:43 < kanzure> for example - http://science.slashdot.org/comments.pl?sid=521078&cid=23060336
07:43 < kanzure> that's an utter failure of reasoning
07:43 < kanzure> http://science.slashdot.org/comments.pl?sid=521078&cid=23060922 might have some sense
07:45 < kanzure> http://science.slashdot.org/comments.pl?sid=521078&cid=23059182 moving intellectual processing into automatic responses and triggering - re: reactions / defense mechanisms last night
07:48 < kanzure> http://science.slashdot.org/comments.pl?sid=521078&cid=23058772 "The idea that physical forces control us is silly unless you believe in dualism, we *are* those physical forces."
07:50 < kanzure> For some reason there is this very large pressure from science fanboys to bash the concept of 'free will', but I like to take the Pascalean approach to it: if you have free will and you don't do everything you can, you screw up. If you do have free will and you do as much as possible, congrats. If you don't, then you've maximized the situation anyway.
07:51 < kanzure> I suspect the confusion is because other people want to use it as a way to assess responsibility and blame and other bullshit
07:51 < kanzure> *If you don't [have it], then you've maximized the situation anyway.
08:05 < kanzure> weird, I was testing out mod_rewrite and found http://heybryan.org/test/
08:06 < kanzure> seems to be an XMLHttpRequest testbed
08:17 < kanzure> wtf was this? http://heybryan.org/wiki
08:18 < kanzure> mv wiki wiki_old
08:49 -!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] has quit ["Leaving."]
10:12 < faceface> gah! what did I miss?
10:12 < faceface> I am compiling a list of protein structures with 'open access' articles.
10:13 < faceface> I'll upload them here soon... http://PDBWiki.org
13:12 < fenn> i havent seen anything worth reading on slashdot in a long time
13:13 < fenn> not in the sense of news at least
13:18  * faceface listens to the radio only
14:43 -!- Splicer [n=p@h98n3c1o261.bredband.skanova.com] has joined #hplusroadmap
17:42 -!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] has joined #hplusroadmap
18:01 < fenn> i've been reading about XML Schema, and decided I hate it
18:01 < kanzure> what about DTD?
18:01 < fenn> instead, i think i'm going to use YAML, and validate the _data structure_ when it's alive in code
18:01 < fenn> instead of validating a dead document
18:02 < fenn> yaml just represents the data structure, there's no built-in validation, so it separates the complexity into more manageable chunks
18:02 < fenn> and, xml is fucking impossible to read
18:02 < kanzure> validate the runtime data struct? huh?
18:03 < fenn> right
18:03 < fenn> its like duck typing
18:03 < kanzure> how does that work? if the data can't be serialized into the struct, then it's invalid?
18:03 < kanzure> hm
18:03 < fenn> well, i'm thinking it would be de-serialized (parsed) automatically, by yaml
18:04 < fenn> so we'd write the individual package metadata in yaml, and the validation rules in (any language that can read yaml)
18:04 < kanzure> would we also write the package file in yaml?
18:04 < fenn> i'm only talking about the package file (so far)
18:05 < fenn> say you had a chair module, that has specialized attributes/units of "comfiness"
18:05 < fenn> the yaml would be something like comfiness: 2.0
18:06 < fenn> this gets parsed into an object with attribute comfiness, like mychair.comfiness
18:06 < fenn> now, if there's no specialized code to use that attribute, nothing happens to it
18:06 < fenn> or, if the code is expecting this attribute to exist, it will complain in the standard way "no attribute"
18:07 < fenn> it took me a while to figure out that yaml only represents a (annotated) data structure
19:04 < kanzure> fenn: http://techshop.ws/
19:58 < kanzure> I need to go read up on yaml, because I still don't understand that explanation
19:58 < kanzure> would we use a verification module that loads up the struct into a runtime instance, or would we wait for validation to occur at the last possible mometn?
19:58 < kanzure> *moment
20:13 -!- Aulere [n=dragon_d@131.229.176.252] has joined #hplusroadmap
20:13 -!- Aulere [n=dragon_d@131.229.176.252] has left #hplusroadmap []
20:20 -!- Aulere [n=dragon_d@131.229.176.252] has joined #hplusroadmap
20:21 < kanzure> Aha.
20:21 < Aulere> :)
20:27 < kanzure> yaml looks carefully written, fenn - http://yaml.org/spec/cvs/current.html
20:41 < kanzure> fenn: would we just write classes in python and then call some yaml libraries to serialize the data directly?
20:42 < kanzure> it looks like we don't have to do any yaml by hand, which might be interesting
20:46 < fenn> yes that would be one way to do it, for example setting up templates
20:46 < fenn> then mortals can go in and edit the yaml
20:47 < fenn> then we read the yaml back in to process dependencies and functionalities
20:47 < kanzure> not bad
20:48 < kanzure> so in that case, what does 'dependencies' refer to?
20:48 < kanzure> there are many types of dependencies that we've been throwing around obviously
20:48 < fenn> dependencies is technological specifications, interfaces and signal types and measured units
20:49 < kanzure> All dependencies throughout the entire project?
20:49 < fenn> like, chair is made of wood
20:49 < kanzure> that sounds like it might work, but I don't want to be too eager 
20:49 < kanzure> oh, so if yaml needs a dependency for one of the serialized structs, it can spit out a specific error
20:50 < kanzure> and then the client can go grab 'em
20:50 < kanzure> ideally the metadata would specify which package has that, yes?
20:50 < fenn> yaml doesnt do anything but turn flat text files into live data structures
20:51 < fenn> something has to come along and 'validate' those structures, by examining them
20:51 < fenn> if it doesn't know what an element in the struct is, it says 'help!' and looks in the mime database for a specialized module
20:52 < kanzure> mime database? or is it now a yaml database?
20:52 < fenn> i dont know exactly what to call that type of program
20:52 < fenn> yaml can store mime information right?
20:52 < fenn> well, it doesnt matter
20:52 < kanzure> ok
20:52 < kanzure> oh, what if it was a yaml querying language
20:52 < fenn> there's a field, common to all package files, that describes the modules needed to process the package dependencies
20:52 < kanzure> would that be a good way to call it?
20:52 < kanzure> probably a list of fields, but yes
20:53 < fenn> what's a yaml querying language?
20:54 < kanzure> sounds like what we just described -- have a database that we can query to get missing pieces of a puzzle, a partially 'self-correcting' (self-validating) code structure, and when it comes across ambiguities it asks for user help (and generates log files, error reports, etc.)
20:54 < kanzure> if it doesn't find something, then its "help!" is basically a query, isn't it?
20:54 < kanzure> "find me something to fix this problem"
20:54 < fenn> so a specialized module would be 'chair' and it would get called when you run across the 'chair:' element while traversing the tree
20:54 < kanzure> sure
20:54 < kanzure> and the chair might require a wood package or something
20:54 < kanzure> and the local user doesn't have wood downloaded
20:55 < fenn> yes, 'help' is a query
20:55 < kanzure> so the yaml says "help!" (as a query) and the local database finds it has metadata that points towards somewhere on the web to download the entire package contents (not just the package file / metadata for the package)
20:55 < fenn> no, yaml isnt doing anything
20:55 < kanzure> hm
20:56 < fenn> lets give the process a name, maybe 'autospec'
20:56 < fenn> autospec is concerned with understanding specifications
20:56 < kanzure> is this runtime autospec, or pre-runtime? (maketime, call it)
20:56 < kanzure> maketime is ambiguous, nevermind
20:56 < fenn> it traverses the tree generated from dead yaml files, examining each element, making sure they are consistent and make sense
20:57 < kanzure> how does it fill in the gaps?
20:57 < kanzure> or identify a gap?
20:57 < fenn> what gaps?
20:57 < kanzure> if you have chair, and it requires wood, you don't have wood installed yet
20:57 < kanzure> shouldn't it be able to identify that situation?
20:57 < fenn> that's later on
20:58 < fenn> that's the package manager's job
20:58 < kanzure> this is just package-file / metadata-file validation?
20:58 < kanzure> not within the context of other packages
20:58 < kanzure> just local, very specific validation
20:58 < fenn> right
20:58 < fenn> just make sure we arent getting bad input, and that we have everything needed to understand the input
20:58 < fenn> obviously you dont know if it's good or bad if you cant read it
20:59 < fenn> by read i mean, have methods that correspond to each element in the specification
21:00 < kanzure> why do we need autospec for the end-user?
21:01 < kanzure> just validate it once, at the repo
21:01 < fenn> to make sure they have software needed for processing the package file they are looking at (or recursively depended upon somehow)
21:01 < fenn> then you require all users to have up to date software installations if they want to be able to read a package
21:01 < kanzure> how would whether or not something passes yaml-syntax tell you if you have the software needed to work with that file format?
21:01 < kanzure> hm
21:01 < fenn> then you get versions like sid etch lenny, etc
21:02 < fenn> if you have methods that understand the package dependencies, that's the software
21:02 < kanzure> I don't think that answers my question
21:03 < kanzure> yaml-synatx validation is not the same thing as checking your MIME-lookup-database-tables to see if you have software for working with such-and-such fileformat.
21:03 < kanzure> (in this case we get to s/MIME/yaml/ since yaml doubles for MIME)
21:03 < fenn> MAML :)
21:04 < kanzure> so
21:04 < kanzure> load up all of the file format data structs at the same time, and then yaml doesn't cry if it comes across something new?
21:04 < kanzure> that could be a lot of overhead
21:04 < fenn> if you dont have software that can do anything with the specification, why bother reading it at all?
21:04 < kanzure> can we make yaml do a lookup in our local database?
21:04 < kanzure> (the MIME db, or as you call it, MAML)
21:05 < fenn> autospec would be the one crying
21:05 < kanzure> autospec retries once yaml fails? (retries after loading up a supposed fix, that is)
21:05 < fenn> no no, yaml just reads text files and spits out code objects
21:06 < fenn> autospec catches those code objects and looks at them, if it doesnt understand what they are it queries the mime db
21:06 < fenn> if the db has nothing, then you try to download new software or give up
21:06 < kanzure> what does understanding consist of
21:07 < fenn> mathematical models, essentially
21:07 < kanzure> giving the name of a program that can work with that file format?
21:07 < kanzure> *that data format
21:07 < fenn> also, relations between specific logical symbols
21:07 < kanzure> what does autospec call to the mime db besides 'lookup this'
21:07 < kanzure> and what does mime db return?
21:07 < fenn> autospec already knows the name of the program that can work with the file format (if it doesnt, it queries the db)
21:08 < kanzure> how would autospec already know? have a lookup table of its own, filled with commonly needed programs?
21:08 < kanzure> *program pointers / paths
21:08 < fenn> the db returns an autospec plugin that can verify whether or not the two packages are compatible
21:08 < kanzure> hm
21:09 < fenn> so, say you have a tire, you want to know if it goes on a car
21:09 < kanzure> the mime db *contents* and the autospec plugins should be the same thing, but mime db has the daemon that should be able to fetch other software and so on
21:09 < kanzure> yes?
21:09 < fenn> you look at the tire spec, you look at the car spec, call some specialized code in either car or tire modules, and it returns a 'yes' or 'no'
21:10 < fenn> or, say you want to make the tire, you run some program on the rubber-molder to see if it can do it
21:11 < kanzure> how would that look like on the shell
21:11 < kanzure> # autospec rubberMolder my.tire.file ?
21:11 < fenn> yep
21:11 < fenn> i guess
21:11 < kanzure> what about my message re: mime db *contents* being the same thing as the autospec plugins?
21:12 < kanzure> they are basically the same thing
21:12 < fenn> the car/tire thing is easier because it's a defined interface
21:12 -!- Splicer [n=p@h98n3c1o261.bredband.skanova.com] has quit []
21:12 < fenn> yes they are the same thing
21:13 < fenn> we can make wrappers so autospec can use external programs
21:13 < fenn> but probably most of it will be new code
21:13 < fenn> you know, stuff like, does a class 4 bolt fit in a class 2 bolt hole
21:14  * kanzure is still digging http://heybryan.org/music/Scenemusic%20-%20live.mp3.m3u
21:14 < kanzure> now, I was originally thinking that mime db would have some daemons to manage its collection or something
21:15 < kanzure> so that it could go out and make a request to a meta-repository somewhere to get new software for a specific file format that it doesn't know how to deal with
21:15 < kanzure> this is why even though they are the same thing, there's still a difference
21:15 < kanzure> or would autospec deal with fetching ?
21:15 < kanzure> would we want it separated? 
21:15 < fenn> i'd rather autospec just look at files and load plugins
21:15 < fenn> not fetch new software
21:16 < kanzure> then the mime db can have a software interface
21:16 < fenn> unix philosophy of small tools that do one thing well
21:16 < kanzure> so that autospec doesn't know if it's getting plugins from the net or locally
21:16 < fenn> yes
21:17 < kanzure> how does it load plugins? yaml parses the file, and then gives us an object that we can work with; do we just do "object matching"? I don't know what that means, frankly
21:18 < fenn> you could look at each of the object's attributes and see if you have methods for them
21:18 < kanzure> http://openbabel.org/wiki/Main_Page Open Babel is a chemical toolbox designed to speak the many languages of chemical data. It's an open, collaborative project allowing anyone to search, convert, analyze, or store data from molecular modeling, chemistry, solid-state materials, biochemistry, or related areas.
21:18 < kanzure> so then every new object must have new attributes?
21:18 < kanzure> I think this is what duck typing is about?
21:18 < fenn> um, i dont understand why you're asking that?
21:18 < fenn> an object is like a box full of stuff
21:19 < kanzure> yes
21:19 < kanzure> and it has attribute variables
21:19 < kanzure> you say match the attributes to the mime db
21:19 < kanzure> to an entry in the db, I mean
21:19 < kanzure> so if I have 5 attributes for car, and 5 attributes for window, is that a conflict?
21:19 < fenn> no, because a car can have windows
21:20 < fenn> uh. hm.
21:20 < fenn> i dont want this to turn into cyc
21:20 < kanzure> we'd have to do unique names for attributes
21:20 < fenn> namespaces
21:20 < kanzure> yes
21:20 < kanzure> or it's all in the same namespace, and the code has weird hash ids for the variables, but with a human readable overlay map
21:21 < fenn> it boils down to the same thing
21:21 < kanzure> so that when you submit code to the database, the variable names are replaced with IDs, but a 'human key/legend' is given down
21:21 < kanzure> *is given
21:21 < kanzure> which way would be simpler?
21:21 < fenn> c++ does that with 'mangling' and i think its silly
21:21 < kanzure> how would we implement namespaces?
21:21 < fenn> by module
21:21 < fenn> car, window
21:21 < fenn> you say object.car.wheel
21:22 < fenn> so the first layer autospec looks at is the top layer (car)
21:22 < kanzure> okay, but this isn't entirely centralized
21:22 < fenn> it loads the car module, then car.wheel makes sense too
21:22 < kanzure> what if I have a package obj.car.wheel and then from another one I have obj.airplane.wheel or something
21:22 < kanzure> and there's overlap, but there's not mean to be overlap between car wheel and airplane wheel
21:22 < kanzure> thus a namespace violation
21:22 < kanzure> conflict.
21:22 < fenn> in the car module, you can inherit from the 'wheel' class which is defined elsewhere (and so does the airplane module)
21:23 < fenn> because a car 'has' wheels
21:23 < kanzure> okay
21:23 < fenn> the car class has an attribute 'wheel' that isnt necessarily defined in the car class code
21:24 < fenn> i dont know the gory details, but it involves function pointers
21:24 < kanzure> I thought autospec would resolve those glory details
21:24 < fenn> no, that's just an object oriented language
21:24 < kanzure> python does it natively, yes?
21:24 < fenn> yes
21:25 < fenn> so, the car module would have a software dependency on the wheel module
21:25 < fenn> resolved by some apt-like system
21:25 < kanzure> but the car would also have a physical dependency on it, in the final project/ implementation
21:26 < kanzure> and this has to be specified too somehow
21:26 < fenn> yeah, and that's the more squishy problem
21:26 < fenn> does a car have 4 wheels? what about one of those three-wheeler thingies
21:26 < fenn> what about a 6 wheel dumptruck
21:26 < fenn> is that a car?
21:26 < kanzure> right, so there would be a variety of classes and objects
21:27 < fenn> i'd rather make the definitions broad, since you're just calling it a car to access the car class's code
21:27 < kanzure> well, you could at the very least create a table or matrix to suggest somebody some axis of definitions, i.e. one variable of adjustment could be the number of wheels
21:27 < kanzure> yep
21:27 < kanzure> and also the (broad) manufacturing code
21:27 < fenn> i mean, it could be a toaster with a v6 engine, not necessarily a car, but you want to use the specifications provided by the car class
21:27 < fenn> maybe automobile is more appropriate
21:28 < fenn> generalize until it doesnt mean anything, eh wot
21:28 < kanzure> well, the idea is that specialized people will come in and add projects
21:28 < kanzure> and then some others can go link to it back and forth on other projects so that more people can be aware of it
21:28 < kanzure> that's all editable, we don't do top-down in general
21:28 < kanzure> it's definitely going to be bottom-up contributions, but that's okay, some of us will work top-down, trying to connect the dots
21:29 < fenn> you dont want a million differnt classes that do the same thing
21:29 < kanzure> thus refactoring, redirection, etc.
21:29 < fenn> someone's gotta come through and refactor, right
21:30 < kanzure> we have to do this *anyway* without skdb
21:30 < fenn> we have to refactor what?
21:30 < kanzure> (I've taken to calling this 'metarepo' actually, for obvious reasons - it's a ghost layer for aggregation of packages with definite specifications, and the metarepo itself can be copied by anybody -- preferably with a solid connection back to our instance)
21:30 < kanzure> hm?
21:30 < kanzure> refactor manufacturing projects
21:30 < kanzure> think about how you do an electrical circuit at the moment
21:31 < kanzure> you refactor and reuse various schematics from before
21:31 < kanzure> no?
21:31 < fenn> sure
21:31 < kanzure> well, this sort of refactoring is just much more useful IMO
21:31 < fenn> sometimes i'll even get lucky and it will do exactly what i want straight from te plans
21:31 < kanzure> has that actually happened?
21:31 < kanzure> it's been a while since I've done electronics
21:31 < fenn> yes, a laminator temperature controller
21:32 < fenn> i didnt actually build it, but i was going to
21:35 < fenn> http://thomaspfeifer.net/laminator_temperatur_regelung.htm
21:36 < fenn> it's very hard to search for electronic circuits
21:37 < kanzure> it's ok, skdb will help fix that - it's a giant aggregation model - one of the aspects of file formats for packages that I will insist on is a module for specifying social network data, like email addresses, mailing lists, specific websites, phone numbers, etc.
21:37 < fenn> upstream contact
21:38 < kanzure> is that a ref to something else, or a terming?
21:38 < fenn> its the debian dterminology
21:38 < kanzure> I've never seen it before.
21:39 < fenn> see, the debian team maintains a copy that closely follows the original package, but has modifications that make everything 'just work'
21:39 < kanzure> huh, that's interesting, so they're really just coding for themselves
21:39 < kanzure> and then release snippits and tidbits in formal releases
21:40 < fenn> usually the person who wrote the original package doesnt want to go through all the trouble of learning how to make a debian package and comply with debian policy like where files go and so on
21:40 < fenn> its more like, they maintain a patch set
21:40 < fenn> sometimes the original author is the debian maintainer
21:41 < fenn> in which case he would either maintain a patch set for debian, or put some code that does those actions during compilation
21:41 < fenn> so, you might have a car designer, but he just knows how to draw pretty pictures
21:42 < fenn> and another guy who knows nothing about car design, but he can assimilate 500 tech projects in a week
21:42 < fenn> well, nothing is stretching it
21:42 < fenn> but there's no requirement for genius
21:43 < kanzure> yes, from what I've seen there are debian evangelicalists (which is not bad at all) that go around trying to help people putting their software into packages
21:44 < fenn> many people dont understand that their projects will be used by others with widely varying tools and situations
21:44 < fenn> so they just write the specifics of what they did
21:44 < fenn> well, that doesnt help 90% of the people out there, the traditional way to deal with it is to re-factor everything by hand
21:44 < fenn> (again, for your specific situation)
21:45 < fenn> so if you dont have a welder, you'd redesign all the parts to be bolted together
21:45 < fenn> or if you cant cast aluminum, you'd make some CNC code to cut it out of a block
21:45 < fenn> or if you dont like PIC chips, you'd re-write the code for AVR
21:46 < fenn> these all change the structure, but the functionality is generally equivalent
21:46 < fenn> i'd like to get the original specification in a form that can be 'compiled' into either of the generally equivalent ways of doing it
21:47 < fenn> maybe it's too much to ask for PIC/AVR
21:47 < kanzure> atm the equivalency is only 'socially known' rather than formally specified such that a file format converter could do the task
21:47 < kanzure> but we're getting there, probably
21:47 < fenn> well, nothing is exactly equivalent
21:47 < kanzure> oh, as for PIC and AVR, sure
21:48 < fenn> i mean, welded joints are usually stronger and more lightweight than bolted joints
21:48 < fenn> but they induce distortion
21:48 < fenn> so you have to say exactly why you're using a welded joint
21:48 < fenn> usually it's just 'because i had a welder and know how to use it'
21:49 < fenn> and then you get into role-playing game stuff, modeling people and their skills
21:50 < fenn> how much they're willing to learn/work vs spend money vs product quality
21:56 < kanzure> guess I need to do some py-yaml
21:57 < fenn> the other thing i figured out today was that 'eval()' is generally evil
21:57 < kanzure> deceptive functionality?
21:58 < kanzure> I remember making excessive use of php's eval() function back in the day
21:58 < fenn> easy to break, allows security holes that can hose your entire system
21:58 < kanzure> I was storing php in mysql 
21:58 < fenn> so we need to sign packages and verify them before being used automatically
21:59 < kanzure> hm
21:59 < kanzure> "YAML represents type information of native data structures with a simple identifier, called a tag. Global tags are URIs and hence globally unique across all applications. The “tag:” URI scheme is recommended for all global YAML tags. In contrast, local tags are specific to a single application. Local tags start with “!”, are not URIs and are not expected to be globally unique. YAML provides a “TAG” directive to make t
21:59 < kanzure> URIs. interesting.
21:59 < kanzure> makes things easy I guess
21:59 < kanzure> it's like the XML's use of DTD
21:59 < fenn> org.yaml.int or something?
21:59 < kanzure> yes
21:59 < kanzure> can we use that to our advantage?
21:59 < fenn> yeah
21:59 < kanzure> giving a hint to the mime db?
21:59 < kanzure> or do we want that to be the direct information for mime db?
22:00 < kanzure> or what?
22:00 < fenn> well, it provides plenty of namespaces
22:01 < fenn> org.skdb.autospec could actually be downloaded from git.autogenix.org?
22:01 < fenn> i'd rather have a layer of redirection available
22:04 < fenn>  the existing YAML libraries come with adequate, but not great, conversion tools for moving between XML and YAML
22:04 < fenn> so the choice isnt set in stone
22:06 < kanzure> sure, we can be that layer
22:06 < kanzure> doesn't add but a handful of ms to processing time anyway
22:06 < fenn> the layer is your db
22:06 < kanzure> right, but autospec doesn't know if mime db is getting the plugins locally or from the net
22:06 < kanzure> so that's the 'layer of redirection'
22:07 < kanzure> yes?
22:07 < fenn> autospec could care less who downloaded the plugins
22:08 < fenn> the redirection is between yaml object type and download url
22:08 < fenn> usually the type and download url would be similar
22:08 < fenn> like org.autogenix.automobile
22:09 < fenn> would point to autogenix.org/v1.23/automobile.deb
22:09 < kanzure> that has a nice, direct mapping to directory structure for the database
22:09 < fenn> or python egg or watever
22:09 < fenn> sure, and then it would install in /usr/share/autogenix/autogenix/automobile
22:10 < fenn> or /usr/share/skdb/autogenix/automobile
22:10 < fenn> depending on exactly what name goes with what code
22:10 < kanzure> autogenix is the client, I think it's just /usr/share/skdb/automobile/ or ~/.skdb/ (whatever) / packages/automobile
22:10 < kanzure> dunno
22:10 < kanzure> hm
22:11 < fenn> then you have a namespace conflict in skdb/
22:11 < kanzure> so we want the code, metadata, software, various files, all separated into different directories?
22:11 < kanzure> or do we want one package to have its own dir?
22:11 < kanzure> and that's final?
22:11 < fenn> um, i dunno
22:11 < fenn> less directories is better usually
22:11 < kanzure> but might mess up the $PATH for the user's os
22:11 < fenn> eh?
22:11 < kanzure> *user's env-variables
22:12 < kanzure> I thought it was good practice to respect the user's path variable
22:12 < fenn> env variables are inherited from te parent shell
22:12 < kanzure> and not go off adding random new files into your own dirs, just follow the /usr/lib/share/ specs for that distro
22:12 < kanzure> right?
22:12 < kanzure> okay
22:12 < fenn> if skdb is an installed program it will inherit the default $path, if the user installs it in their home dir and runs it as the user, it will use the user's $path
22:14 < fenn> i've learned from emc2 that it's nice to have various development versions contained in a single directory, and then you can cd into that directory and run a script that changes your env variables to use that copy as if it were installed
22:14 < fenn> each contained in a single directory of their own, like ~/skdb.stable ~/skdb.testing
22:15 < fenn> but for an installed program on an end-user's machine, it doesnt matter so much, you can follow the distro's guidelines
22:17 < fenn> oops: http://autogen.sourceforge.net/
22:18 < kanzure> Hm.
22:18 < fenn> that's a near-miss, and probably where i got the name from
22:18 < kanzure> :)
22:19 < fenn> its very much what i'm trying to do though
22:19 < fenn> as little source code as possible
22:20 < kanzure> "A common example where this would be useful is in creating and maintaining the code required for processing program options. Processing options requires multiple constructs to be maintained in parallel in different places in your program"
22:20 < kanzure> "options maintenance" - excellent
22:20 -!- Jarocks [n=Jarocks@c-68-80-219-150.hsd1.pa.comcast.net] has joined #hplusroadmap
22:20 < kanzure> Hi Jarocks.
22:20 < Jarocks> hi
22:21 < kanzure> fenn: I wonder if their options maintenance is extensible to the point where it would work with our yaml system.
22:21 < fenn> Jarocks: we're talking about an open-source engineering expert system
22:21 < Jarocks> Okay
22:21 < fenn> that we're making
22:21 < kanzure> I found Jarocks in #biology, he does microbiology. 
22:22 < kanzure> so that means self-replicating (bio) machines
22:22 < fenn> it may or may not be extendable to biotech, hard to say at this point
22:22 < fenn> and my relative ignorance of modern bioinformatics
22:22 < kanzure> faceface can help with that
22:22 < fenn> i've used clustal if that helps :P
22:23 < fenn> kanzure: how are you envisioning options maintenance being used?
22:24 < fenn> i think they mean stuff like, myprogram --verbose --yeah-i-really-mean-it
22:24 < Jarocks> What language is this being coded in?
22:24 < kanzure> fenn: Wasn't that the idea of the 'file format io spec' metadata for programs? So that you can say "This program allows xyz input, as this parameter."
22:24 < fenn> Jarocks: python, shell, yaml (more of a data format)
22:24 < kanzure> Jarocks: Anything that ends up working. Right now it looks like python. We're doing some intense specifications.
22:24 < Jarocks> Ahh
22:24 < kanzure> Jarocks: There's an introduction to this project, sort of, at http://heybryan.org/mediawiki/index.php/Skdb
22:25 < kanzure> http://heybryan.org/mediawiki/index.php/Self-replication
22:25 < kanzure> and a more lucid, but varying attempt at http://heybryan.org/recursion.html
22:25 < fenn> nice block of text on that last link :)
22:25 < kanzure> fenn - I may have added some text to http://heybryan.org/mediawiki/index.php/2008-04-14
22:25 < kanzure> heh
22:25 < fenn> a textoid plaque
22:26 < kanzure> I ... can't help it. :)
22:27 < fenn> autoengix :P
22:27 < Jarocks> Wow I'm way over my head here
22:27 < Jarocks> Lol
22:28 < fenn> not quite as memorable
22:28 < kanzure> so are we
22:28 < Jarocks> So this is like an open source labtable?
22:28 < Jarocks> Sort of?
22:28 < kanzure> That's a good way to put it.
22:28 < kanzure> Have you heard of debian?
22:28 < Jarocks> yes
22:28 < kanzure> Then have you heard of apt?
22:28 < fenn> Jarocks: imagine you have a santa claus machine, that can spit out almost anything
22:28 < Jarocks> I prefer ubuntu but go on
22:28 < fenn> this is the software that the machine uses to build the tools to build the tools..
22:28 < fenn> and to know which tools it needs to build for the end product
22:28 < Jarocks> as in sudo-apt get
22:28 < Jarocks> ?
22:28 < fenn> yes
22:28 < kanzure> Jarocks: APT for projects.
22:29 < kanzure> not just software =)
22:29 < fenn> apt goes and fetches all the programs and libraries you need to run software
22:29 < fenn> autogenix builds all the tools you need to build (stuff)
22:29 < fenn> or tells you what tools you need to get
22:29 < Jarocks> ahh
22:30 < fenn> or determines whether its possible to build X given Y tools and materials
22:30 < Jarocks> Ahh
22:30 < Jarocks> Related to biology
22:30 < Jarocks> okay
22:30 < Jarocks> I got it
22:30 < Jarocks> sort of
22:30 < fenn> yes, actually it might be easier with biology in some respects (patent issues aside)
22:31 < Jarocks> Ahh
22:31 < Jarocks> okay
22:31 < fenn> a single cell is way more sophisticated and integrated than human technology is right now
22:32 < Jarocks> yep
22:32 < fenn> but, its not easily programmable. there's no documentation
22:33 < Jarocks> For cells?
22:33 < Jarocks> or desbianm
22:33 < fenn> and you're limited to a certain range of temperature, material types, and lacking precision in large structures (except where it's a result of emergent properties)
22:33 < fenn> cells are not easily programmable
22:33 < fenn> you cant say, 'make me a toaster'
22:34 < kanzure> Neat, Brett just came up to ask me about the synthesis of ethylene from methanic atmospheres.
22:34 < fenn> hmm.. heavy metal catalysis? how would one go about that?
22:34 < Jarocks> Yes
22:34 < kanzure> I'd like to say make me a toaster - that's what biobricks is supposedly about, but they ignore these stuff.
22:34 < kanzure> fenn: I was thinking it's a simple organic chem reaction method
22:35 < fenn> why does he want to make ethylene
22:35 < kanzure> methane -> ethylene is the addition of a carbon or something
22:35 < fenn> removal of hydrogen
22:35 < kanzure> now that he's quit his job, he just sits around all day doing nothing, supposedly he's writing a scifi manuscript, but from what I've seen him writing previously, eh
22:36 < Jarocks> Do you have PHDs
22:36 < fenn> (ch4)2 -> (ch2=ch2) + h2
22:36 < kanzure> Jarocks: http://heybryan.org/ to see who I am.
22:36 < kanzure> fenn: s/Brett/dad/
22:36 < Jarocks> What do u mean by s/ 
22:37 < kanzure> it's a perl regular expression meaning replace
22:37 < Jarocks> Ahh
22:37 < fenn> it's sed syntax actually :)
22:37 < kanzure> really
22:37 < fenn> replace Brett with dad
22:37 < kanzure> s/$1/$2/ means replace occurences of $1 with $2
22:37 < Jarocks> I've never been great with regular expression
22:37 < fenn> yeah, in sed it's s/\1/\2/ i think
22:37 < kanzure> Me either ;)
22:37 < Jarocks> have always relied heavily on tutorials for it
22:37 < fenn> well, no it's actually s/(stuff)/\1/
22:37 < kanzure> Jarocks: yeah, so we don't have PhDs.
22:37 < fenn> would do precisely nothing
22:38 < fenn> BS piled higher and deeper
22:38 < fenn> i'm a level one academic
22:39 < kanzure> fenn: def?
22:39 < Jarocks> I'm still in AP bio
22:39 < Jarocks> Taking Honors Micro also
22:39 < kanzure> Jarocks: Yeah, I took AP bio last year
22:39 < Jarocks> Gonna take Ap Chem next year 
22:40 < kanzure> http://heybryan.org/school/Biology/notes/output.html are all of my notes from last year
22:40 < kanzure> (on bio)
22:40 < fenn> kanzure: BS == bachelor of science, is your first piece of paper
22:40 < kanzure> that's level one?
22:40 < fenn> then you get your MS == master's degree
22:40 < kanzure> I thought level one would be the associate 
22:40 < kanzure> *associate's degree
22:40 < kanzure> or something
22:41 < fenn> i forget what the order is
22:41 < kanzure> Associate's, BS, MS, PhD/MD, then some other random stuff throughout there to make you 'certified' like the GE for engineers and so on
22:41 < fenn> somewhere in there, there's associate, doctorate, professor, emeritus(?)
22:41 < Jarocks> Do you know anything about the mitosome beyond what wikipedia has
22:41 < kanzure> emeritus is a degree?
22:41 < fenn> well, its more like putting you out to pasture
22:41 < kanzure> Hm.
22:42 < fenn> a mitosome? whats that, like a mitochondria?
22:42 < fenn> mitochondrium
22:42 < kanzure> new organelle, fenn
22:42 < kanzure> http://en.wikipedia.org/wiki/Mitosome
22:42 < kanzure> Unlike mitochondria, mitosomes do not have genes within them. The genes for mitosomal components are contained in the nuclear genome.[1] An early report suggested the presence of DNA in this organelle,[7] but more recent research has shown this not to be the case.[8]
22:43 < kanzure> huh
22:43 < kanzure> that's what http://sens.org/ is looking for
22:43 < kanzure> if the mitosomes have the same functionality as mitochondria, and the genes are in the nuclear genome, then that's exactly it
22:44 < fenn> why do they need this?
22:44 < Jarocks> okay
22:44 < fenn> i thought mitochondria didnt age
22:44 < Jarocks> It doesn't age, It ages you :P
22:44 < fenn> bah
22:44 < Jarocks> Well it ages also
22:45 < Jarocks> They do
22:45 < Jarocks> Indirectly
22:45 < Jarocks> Along with other factory
22:45 < Jarocks> *factors
22:45 < fenn> they dont age any more than bacteria age
22:45 -!- Jarocks [n=Jarocks@c-68-80-219-150.hsd1.pa.comcast.net] has quit [Read error: 104 (Connection reset by peer)]
22:46 -!- Jarocks [n=Jarocks@c-68-80-219-150.hsd1.pa.comcast.net] has joined #hplusroadmap
22:46 < Jarocks> lost connection for a bit
22:48  * kanzure was interrupted again
22:48 < kanzure> fenn: mitochondria cause problems related to aging apparently
22:48 < kanzure> MitoSENS explains
22:49 < kanzure> http://heybryan.org/transhuman/kanzure_questions.html
22:49 < Jarocks> The free-radicals
22:49 < kanzure> search for 'mito'
22:49 < kanzure> Mitochondrial mutations oxidize circulating material and introduce toxins into mitochondrially healthy cells elsewhere. How does oxidized material contribute to senescence?
22:49 < kanzure> [74] de Grey, A.D.N.J. (2003) in Genetics of mitochondrial diseases (Holt, I.J., Ed.), Oxford University
22:49 < kanzure> Muscles are dependent on many thousands of mitochondria per myocyte. How can we localize PMRS inhibitors so as to not cause unintended muscle damage?
22:49 < kanzure> mitochondriopathies
22:49 < fenn> circulating material?
22:50 < kanzure> educe the amount of NADH available to mitochondria (NADH takes electrons away, remember), via adding “intracellular electron acceptors.”
22:50 < kanzure> wild-type mtDNA
22:50 < kanzure> anyway, lots of notes on that page
22:51 < fenn> hayflick limit is interesting
22:51 < kanzure> dunno how that's corrected for in spermatogenesis/embryogenesis
22:51 < fenn> i dont think anyone does (last i heard)
22:52 < Jarocks> Is that related to the whole linear genome shortening of the dna thing
22:52 < fenn> yes, telomeres etc
22:52 < Jarocks> the hayflick limit not apply to prokaryotes
22:52 < Jarocks> *does
22:52 < fenn> or cancer cells
22:53 < kanzure> Jarocks: was that +does or [-not +does] ?
22:53 < Jarocks> +does
22:53 < Jarocks> Does the hayflick limit not apply to prokaryotes?
22:54 < fenn> it does not, otherwise they would all be dead by now
22:54 < Jarocks> Lol
22:54 < kanzure> serves them right
22:54 < fenn> stupid prokaryotes
22:54  * kanzure kicks a few prokaries
22:54 < Jarocks> Doesn't apply to cancer >_<
22:55 < Jarocks> Stupid telemerease
22:55 < kanzure> killed Dolly, and anything that killed Dolly must be bad
22:55 < Jarocks> (not that its stupid, only in the case of cancer)
22:55 < Jarocks> Yep
22:55 < fenn> bwaaaah
22:56 < fenn> (sad sheep noise)
22:56 < kanzure> fenn: you need to work on it
22:56 < kanzure> I mistook it for your mother
22:56 < kanzure> not bad, eh?
22:56 < fenn> i dont get it
22:57 < fenn> are you saying i was cloned from a sheep?
22:57 < kanzure> the idea is to one-up each other with your-mom jokes, but it doesn't matter since I always lose
22:57 < kanzure> my mom's a whore
22:57 < kanzure> a stripper, I mean
22:57 < Jarocks> lol
22:57 < kanzure> http://lockhartwoodworks.com/
22:57 < Jarocks> Idk people don't do your mom jokes much anymore
22:57 < kanzure> I see.
22:58  * kanzure gets back to work
22:58 < Jarocks> I really wanna make protobionts again, was fun yet pointless
22:58 < kanzure> ?
22:58 < Jarocks> Idk we made protobionts in ap bio 
22:59 < Jarocks> it was fun
22:59 < Jarocks> yet they are unsuprisingly boring
23:00 < Jarocks> Idk I like the whole origin of life thingy
23:00 < Jarocks> Abiogenisis
23:00 < fenn> people always assume it happened on earth though
23:01 < fenn> even though life appeared unreasonably soon after the planet surface cooled off
23:01 < Jarocks> Well if not on Earth
23:01 < Jarocks> Then Mars is the next best canidate
23:01 < fenn> bah
23:01 < kanzure> shit, who told Wheeler to die
23:01 < fenn> there's the whole rest of the universe out there
23:02 < kanzure> "Eminent physicist John Archibald Wheeler has died from pneumonia at the age of 96. The coiner of the terms 'black hole' and 'wormhole,' Wheeler popularized the study of general relativity, and advised a distinguished list of graduate students including 
23:02 < kanzure> Kip Thorne and Richard Feynman. Other work included a collaboration with Niels Bohr to develop the 'liquid drop' model of nuclear fission. 
23:02 < kanzure> Max Tegmark, a cosmologist at the Massachusetts Institute of Technology, said of Dr. Wheeler, 'For me, he was the last Titan, the only physics superhero still standing.'"
23:02 < kanzure> Max Tegmark does some good mathematics worth looking into, re: Zindell and Egan.
23:03 < fenn> well, they're all dead, maybe we'll get some new physics now
23:03 < Jarocks> well most people need to get over the fact that life probobly isn't as special as we would all like to think it
23:03 < fenn> life is quite special, why do you think it's not?
23:04 < Jarocks> It is
23:04 < Jarocks> But if panspermia is true
23:04 < kanzure> hm, Weinberg is still alive and local
23:04 < Jarocks> Then life may be a universal norm
23:04 < kanzure> So what if panspermia is true?
23:04 < fenn> then earth isnt quite so special (but it's all we got)
23:05 < fenn> also, i bet there arent any butterflies or mountain lions anywhere else in the universe
23:05 < Jarocks> Lol
23:06 < Jarocks> True
23:06 < kanzure> ah, then there's Dyson and Gell-Mann, and Wolfram, some of them are still alive I guess (Wolfram just hung out with those guys, not necessarily such a biggie, not yet)
23:06 < fenn> well, maybe
23:06 < fenn> hard to say really
23:06  * fenn kicks wifi router
23:06 < Jarocks> lol
23:06 < Jarocks> Mines not close enoug
23:06 < Jarocks> *enough
23:06 < fenn> mine's too far away
23:07 < Jarocks> So are they doing anything about gravity
23:07 < fenn> i have to ping it in order for my ssh session to update sometimes
23:07 < Jarocks> Now that these people died
23:07 < Jarocks> Because I still am not buying into that whole dark matter stuff
23:07 < fenn> well its quite early, i mean, only like 8 hours ago or so
23:07 < Jarocks> I think we got gravity wrong: again
23:08 < fenn> give the man some respect before vulturing into his property!
23:08 < Jarocks> lol
23:08 < kanzure> re: gravity, see John Baez and Lee Smolin
23:10 < Jarocks> ooh nanobacteria
23:10 < kanzure> anybody remember 'surfer-dude physicist', A. Garret Lisi?
23:11 < Jarocks> nope
23:12 < kanzure> meh
23:13 < Jarocks> Proboblu
23:13 < Jarocks> *probobly
23:13 < Jarocks> yeah
23:14 < Jarocks> I think so
23:14 < fenn> nanobacteria could be biological but not self-replicating
23:16 < Jarocks> They have shown they can reproduce
23:17 < Jarocks> theres i think even a pic of one dividing
23:18 < fenn> nanobaclabs.com sets off my quack detector
23:18 < fenn> er, investment-scam detector
23:19 < kanzure> if it looks like a duck, and quacks like a duck, we have at least to consider the possibility that we have a small aquatic bird of the family anatidae on our hands.
23:19 < Jarocks> New superweapon possably
23:20 < fenn> interesting that wickramasinghe (panspermia guy) is an author on one of their papers
23:20 < Jarocks> Heh the superantigen would actually be the best canidate for a bio weapon
23:20 < kanzure> fenn: anybody suggesting panspermia and not doing active astrophysics spectra research, or not doing nucleic acid auto'genesis' research, is a quack
23:20 < kanzure> there's absolutely no reason to dive into those sorts of experiments these days
23:21 < kanzure> uh
23:21 < kanzure> *to not dive into those experiments
23:21 < fenn> um, they already did spectra
23:21 < kanzure> neat
23:21 < fenn> result: bacteria are damn near everywhere
23:21 < fenn> then there's the stardust probe results
23:21 < fenn> (bacteria fragments)
23:22 < fenn> wickramasinghe was trying to isolate bacteria from the stratosphere
23:22 < kanzure> did he succeed?
23:22 < Jarocks> Probobly
23:22 < fenn> dunno, i think they were either hard to culture or pretty normal bacteria
23:23 < fenn> doesn't say much either way
23:23 < Jarocks> These nanobes look awsome
23:23 < fenn> they did isotope analysis of the bacteria, i forget what the results were
23:23 < Jarocks> smaller than 300 nm!
23:23 < Jarocks> Possably the most primative life
23:23 < Jarocks> ?
23:24 < fenn> http://www.panspermia.org/whatsne32.htm#040331
23:24 < fenn> i talked with brig klyce at the alife conference, he seemed like a normal scientist
23:25 < kanzure> are you on the alife mailing list?
23:25 < fenn> no
23:25 < Jarocks> Can you find out what the deal is with the nannobacteria
23:25 < Jarocks> Apparently theres evidence to support them
23:25 < kanzure> http://lists.idyll.org/listinfo/alife-announce
23:25 < Jarocks> and other stuff
23:25 < Jarocks> confusing
23:27 < kanzure> it's mostly postdoc position and conf announcements
23:27 < fenn> it's the -announce list
23:28 < fenn> i think scientists mostly talk one-on-one or write letters to journals
23:28 < fenn> or go to conferences
23:29 < fenn> it's changing of course, but they're not totally decentralized like we are
23:29 < kanzure> I have a file in my ~/cache/ that talks about changing the way sci is done to be more like programmer hackfests
23:30 < kanzure> for more rapid communication, especially over the net and in person for 48 hour periods of intense work
23:30 < kanzure> voluntary cramtime
23:30 < fenn> yeah its pretty stupid that people spend 2 months preparing a speech at a conference
23:30 < kanzure> yikes
23:30 < fenn> well, maybe not that much
23:30 < kanzure> they should do it like perl lightning talks
23:30 < fenn> depends on the conference
23:30 < kanzure> 5 minute talks, stand up, get your point across, sit back down, bitch
23:31 < fenn> hah
23:31 < Jarocks> God I've been searching google, theres nothing
23:31 < Jarocks> Nothing usefull
23:31 < kanzure> Have you ever considered that Google *is* god?
23:31 < fenn> google, i've been searching, there's nothing!
23:31 < fenn> google why have you forsaken me
23:31 < kanzure> "file a bug report"
23:32 < Jarocks> lol
23:33 < Jarocks> I mean nothing usefull
23:33 < fenn> sounds like you need AutoScholar
23:33 < Jarocks> as into sheding the light onto what the hell nanobacterium
23:33 < kanzure> indeed!
23:33 < Jarocks> are
23:33 < fenn> for only 4 million dollars, you can have lightning-fast access to any scientific publication on the planet!
23:33 < kanzure> http://heybryan.org/projects/autoscholar/
23:33 < Jarocks> Lol
23:34 < fenn> this might be a good place to start, if you're really interested (see why i hate accademia) http://www.nanobaclabs.com/content/scientific-publications.htm
23:34 < kanzure> fenn: would be nice to see AutoScholar integration with autogenix (not necessarily skdb)
23:34 < fenn> meh
23:35 < kanzure> would be an interface package perhaps
23:35 < fenn> i dont expect a computer to get anything useful from natural language descriptions
23:35 < kanzure> sure
23:35 < kanzure> but with BibTex being passed around, some good could come of it
23:35 < kanzure> especially if BibTeX is given reference-links to directly get the file
23:35 < kanzure> with open access sci papers, shuffling bits and bytes becomes easier
23:35 < kanzure> in fact, there's no reason to not just pass along a depth-tree of history references
23:36 < fenn> right
23:36 < kanzure> and then if you want the full paper and all references, go to a watering hole somewhere (a packaging/aggregator server for papers)
23:36 < fenn> well, there's no reason not to pass along the whole paper really
23:36 < kanzure> right
23:36 < fenn> anyway..
23:36 < kanzure> and then some references too while we're at it
23:36 < kanzure> it's not much space
23:36 < kanzure> unless it's image-only format
23:37 < kanzure> in which case somebody needs to be shot
23:37 < fenn> i will slay anyone who releases a paper in image format
23:37 < Jarocks> Well you've lost me here
23:37 < Jarocks> lol
23:37 < fenn> its 2008 there's no excuse
23:37 < kanzure> Jarocks: some scientists release papers about their studies, but they do it as an image
23:37 < kanzure> think of it as taking a picture of your paper
23:37 < kanzure> and then uploading the picture
23:37 < kanzure> instead of the plain text
23:37 < Jarocks> IC
23:38 < Jarocks> B/C just putting up a plaintext version would be too easy?
23:38 < Jarocks> What about some sort of text recognition system?
23:38 < kanzure> tried that
23:38 < kanzure> see http://heybryan.org/projects/autoscholar/
23:38 < fenn> because they think the pdf is prettier, and are using broken software
23:38 < kanzure> OCR software really sucks with image-based PDFs anyway
23:38 < kanzure> because normally the paper that they pictured was wrinkly or something
23:38 < fenn> not enough resolution?
23:38 < kanzure> dunno
23:39 < kanzure> that too
23:39 < kanzure> lots of problems
23:39 < fenn> ocr needs lots of resolution still
23:39 < fenn> reading is a very involved process
23:39 < kanzure> yes
23:39 < fenn> you practically need something like cyc to disambiguate
23:40 < fenn> in which case you can just blank out entire sentences and get the gist
23:40 < kanzure> plus some magic =)
23:40 < kanzure> grammatoscope
23:40 < fenn> yeah, check out geoff hinton, that's magic
23:40 < kanzure> zoomable grammars and expandable texts
23:40 < fenn> the ANN stuff
23:40 < kanzure> not for zoomability, that's just "assisted authoring" plus some heavy-duty XMLing
23:41 < fenn> heh
23:41 < kanzure> Japanese cell phones do intense assisted authoring, last I heard
23:41 < kanzure> never tried it
23:41 < fenn> would you like the abstract? or an infinite amount of text? 
23:41 < kanzure> don't tempt me?
23:42 < fenn> i'll take the abstract, thanks
23:42 < fenn> if only they had pictures in abstracts
23:42 < Jarocks> lol
23:42 < fenn> dont lol me, mister
23:42 < kanzure> more SVG for pics too, please
23:42 < kanzure> but photographs are okay
23:42 < fenn> svg for representative art
23:43 -!- Jarocks [n=Jarocks@c-68-80-219-150.hsd1.pa.comcast.net] has left #hplusroadmap []
23:43 < kanzure> He wasn't that cool anyway.
23:43 < fenn> for MEMS/carbon nanotube stuff, electron micrographs are really helpful
23:43 -!- Jarocks [n=Jarocks@c-68-80-219-150.hsd1.pa.comcast.net] has joined #hplusroadmap
23:44 < Jarocks> Well I Have to go
23:44 < Jarocks> Cya
23:44 < fenn> tata
23:44 -!- Jarocks [n=Jarocks@c-68-80-219-150.hsd1.pa.comcast.net] has left #hplusroadmap []
23:44 < fenn> tata box inhibitor!
23:46 < kanzure> 'night
23:46 -!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] has quit ["Leaving."]
23:59 -!- Aulere [n=dragon_d@131.229.176.252] has quit []